CN101068663A - Legged mobile robot and its control program - Google Patents

Abstract

The invention provides a legged mobile robot and control program. In a legged mobile robot, a pivoting motion of a foot sole (22) relative to a leg body is controlled so that, from an intermediate time point in a period of departure of a leg body from a floor to a starting time point of a period of landing of the leg body on the floor, the angle (Theta) of inclination of the foot sole (22) of the leg body relative to the floor surface gradually approaches zero. This eases impact to the foot sole of the leg body at the time of landing on the floor and prevents a slip and spin of the foot sole, thereby enabling stable walking or running.

Description

Legged mobile robot and its control program

technical field

The invention relates to a legged mobile robot and its control program.

Background technique

As is well known, in the past, such a proposal has been proposed, that is, in order to alleviate the impact of landing on the robot when the legged mobile robot goes up and down stairs, etc., a technical method for stabilizing the motion is adopted (see, for example, JP-A-5-318342). .

However, when the robot runs by repeatedly changing the aerial period when all legs are floating from the ground and the landing period when any one of the front ends of the legs is in contact with the ground, the impact of the robot is particularly large when it lands. . In addition, due to reasons such as the moving speed of the robot in the air stage and the angular velocity of the yawing axis are too large, the robot may slip or spin at the soles of its feet when it lands.

Contents of the invention

The present invention aims to provide a legged mobile robot capable of stably walking or running and a control program thereof, which can alleviate the impact when the legs land on the ground, and can avoid slipping or spinning on the soles of the legs. .

In order to solve the above-mentioned problems, the legged mobile robot according to the first aspect of the present invention drives the plurality of legs connected to the base to repeatedly switch the ground reaction force acting on any one of the feet of the plurality of legs. The ground stage, and the air stage in which there is no ground reaction force on the feet of all legs, while moving, it is characterized in that when the air stage is changed to the ground stage, the feet of the legs that are scheduled to land on the ground are opposed The inclination angle on the ground gradually changes, and when the leg touches the ground, the contact surface of the foot is parallel to the ground, thus driving the leg.

According to the legged mobile robot of the first aspect of the present invention, the angle of inclination of the feet of the legs relative to the ground is gradually approached to 0 from the middle moment of the leg's lift-off period to the beginning of the landing period, and thus controlled Leg movements. According to this, since the ground area of the feet (contact surface) of the legs after transitioning from the ground-off phase to the ground phase is relatively large, the impact of landing on the ground can be distributed to the soles of the feet in a wide range, thereby reducing the impact on the robot. shock. In addition, since the friction between the foot (contact surface) and the ground is relatively large, even if the moving speed and the angular velocity of the yaw of the robot before the leg body touches the ground are large, slipping and spinning can be prevented by this friction.

Therefore, the robot of the present invention can alleviate the impact when the foot of the leg body touches the ground, and avoid slipping and spinning of the foot, so that it can walk or run stably.

In addition, the legged mobile robot according to the first aspect of the present invention is characterized in that before the leg is lifted from the ground, the rear end of the foot is moved from the The ground gradually moves away, thus driving the leg body.

According to the legged mobile robot of the present invention, the movement of the legs is controlled in such a manner that the tip of the foot (toe) kicks the ground. According to this, the propulsion force of the robot is enhanced. On the other hand, since the slipping and spinning of the soles of the robot can be prevented when the robot is on the ground as described above, it is possible to make the robot move at high speed while stabilizing its motion.

In addition, the legged mobile robot according to the first aspect of the present invention is characterized in that, from the middle time to the end time of the leg lift-off period, the front end of the foot is raised from the state with the rear end of the foot as a reference. Gradually change to the same height, so drive the leg body.

According to the legged mobile robot of the present invention, from the middle moment to the end moment of the lift-off period, the foot is transformed from the posture of the toes up with respect to the ground to a posture close to being parallel to the ground, so that the foot can be The ground area of the part (contact surface) is ensured to the extent that the slipping of the robot, etc. can be prevented as described above.

In addition, the legged mobile robot according to the first aspect of the present invention is characterized in that the front end of the foot is lowered from the state of the rear end of the foot from the start time to the middle time of the leg lift-off period. Gradually change to the same height, and then gradually change to a higher state, so drive the leg.

According to the legged mobile robot of the present invention, from the beginning moment to the middle moment of the lift-off period, the foot can be changed from the heel-up posture relative to the ground to the toe-up posture, and then be close to The posture of the heel upturned, and then ensure the ground area of the foot (contact surface) to the extent that the slipping of the robot, etc., can be prevented as described above.

In order to solve the above-mentioned problems, a legged mobile robot according to a second aspect of the present invention has an upper body and a plurality of legs extending downward from the upper body, and is connected with a foot that is rotatable with respect to each leg. It is characterized in that it is provided with: a foot inclination angle measuring mechanism, which measures the inclination angle of the foot relative to the ground; and a foot movement control mechanism, which starts from From the middle moment of the ground-off period of the leg body to the beginning moment of the landing period, the angle of inclination of the foot of the leg measured by the foot inclination angle measuring mechanism gradually approaches 0, so that the angle of inclination of the foot relative to the ground is gradually approached. The turning action of the leg.

According to the legged mobile robot of the second aspect of the present invention, the angle of inclination of the feet (soles) of the legs relative to the ground is gradually approached to 0 from the middle time of the leg's lift-off period to the start time of the landing period. , thus controlling the rotation of the foot relative to the leg. Accordingly, since the grounding area of the feet (soles) of the legs after transitioning from the ground-off period to the landing period is relatively large, the impact when landing can be widely distributed to the soles of the feet, thereby alleviating the impact on the robot. shock. In addition, since the friction between the soles of the feet and the ground is relatively large, even if the moving speed and the angular velocity of the yaw of the robot before the legs land on the ground are large, slipping and spinning can be prevented by this friction.

Therefore, the robot of the present invention can alleviate the impact when the feet of the legs land on the ground, and can also avoid slipping and spinning of the feet, so that walking or running can be performed stably.

In addition, the legged mobile robot according to the second aspect of the present invention is characterized in that the foot motion control mechanism controls the rotation motion of the foot relative to the leg body in such a manner that, before the ground lift period, In the state where the leg body is still on the ground with the front end of the foot, the inclination angle of the foot relative to the ground measured by the foot inclination angle measuring mechanism is toward the positive side where the rear end of the foot is farther away from the ground than the front end increase.

According to the legged mobile robot of the present invention, the rotation motion of the foot relative to the leg body is controlled such that the front end (tip) of the foot kicks the ground. According to this, the propulsion force of the robot is enhanced. On the other hand, since the slipping and spinning of the soles of the robot can be prevented when the robot is on the ground as described above, it is possible to make the robot move at high speed while stabilizing its motion.

In addition, the legged mobile robot according to the second aspect of the present invention is characterized in that the foot movement control means controls the rotational movement of the feet relative to the legs, that is, from the middle of the ground-off period of the legs When the time arrives at the start of the landing phase, the angle of inclination of the foot relative to the ground measured by the foot inclination angle measuring mechanism gradually decreases to 0 from the angle at which the front end of the foot is farther away from the negative side of the ground than the rear end.

According to the legged mobile robot of the present invention, from the middle moment of the lift-off period to the start moment of the landing period, the feet are changed from the posture of the toes up with respect to the ground to a posture close to being parallel to the ground, so that The ground area of the soles of the feet is ensured to such an extent that slipping of the robot, etc. can be prevented as described above.

In addition, the legged mobile robot according to the second aspect of the present invention is characterized in that the foot motion control means controls the rotation motion of the foot relative to the leg body in such a manner that, from the start of the leg body's lift-off period, When the time comes to the beginning of the landing period, the inclination angle of the foot relative to the ground measured by the foot inclination angle measuring mechanism gradually increases toward the positive side and then gradually decreases, and then moves toward the front end of the foot compared to the rear end. The negative side away from the ground gradually increases, and then gradually decreases to 0.

According to the legged mobile robot of the present invention, from the start moment of the lift-off period to the start moment of the landing period, the feet are changed from a heel-up posture relative to the ground to a toe-up posture, and then made to Close to the posture parallel to the ground, it is possible to secure the ground area of the soles of the feet to such an extent that the slipping of the robot, etc. can be prevented as described above.

Furthermore, the legged mobile robot according to the second aspect of the present invention is characterized in that it moves while all legs are off the ground in the air phase.

According to the legged mobile robot of the present invention, when the leg body is on the ground from the air, although the leg body is in a state where the other leg body is on the ground and the situation of the leg body is on the ground, the landing impact of the leg body is relatively large, but by As mentioned above, a large landing area can be ensured to reduce the impact at the time of landing.

In order to solve the above-mentioned problems, the control program according to the first aspect of the present invention is a program that provides, to a computer mounted on the robot, a function for controlling a legged mobile robot that drives A plurality of legs connected to the base body, while repeatedly switching the ground reaction force acting on any one of the plurality of legs, and the ground reaction force acting on the feet of all the legs. The control program is characterized in that the computer mounted on the robot is provided with the following function: when the transition from the aerial phase to the landing phase, the legs that are scheduled to land The angle of inclination of the feet relative to the ground gradually changes, and when the legs land on the ground, the contact surface of the feet is parallel to the ground, so as to control the function of the movements of the legs of the robot.

According to the control program of the first aspect of the present invention, the following function is provided to the computer mounted on the robot, that is, the function can alleviate the impact when the foot of the leg body touches the ground, and can also avoid The foot slips and spins so that it can walk or run stably, thus controlling the function of the robot.

In addition, the control program according to the first aspect of the present invention is characterized in that the computer mounted on the robot is provided with the function of: before the leg leaves the ground, the foot of the leg Under the condition that the front end of the foot is still on the ground, the rear end of the foot is gradually separated from the ground, so as to control the function of the movement of the legs of the robot.

In addition, the control program according to the first aspect of the present invention is characterized in that the computer mounted on the robot is provided with the following function: from the middle time to the end time of the leg lift-off period, The function of controlling the movement of the legs of the robot by making the front end of the foot gradually change from a higher state with respect to the rear end of the foot to the same height.

In addition, the control program according to the first aspect of the present invention is characterized in that the computer mounted on the robot is provided with the following function: from the start time to the middle time of the leg lift-off period, The function of controlling the movement of the legs of the robot by making the front end of the foot gradually change from a lower state based on the rear end of the foot to the same height, and then gradually become higher.

In order to solve the above-mentioned problems, the control program according to the second aspect of the present invention is a program that provides, to a computer mounted on the robot, a function for controlling a legged mobile robot having the following functions: body, and a plurality of legs extending downward from the upper body, and are moved by the movements of the legs accompanying the lifting and landing of the feet that can rotate relative to the legs. The control program It is characterized in that the following foot inclination angle measurement function and foot motion control function are provided to the computer mounted on the robot, and the foot inclination angle measurement function is to measure the inclination angle of the foot relative to the ground The foot movement control function is: from the middle moment of the leg's ground-off period to the start moment of the landing period, so that the inclination angle of the foot of the leg relative to the ground measured by the inclination angle measurement function of the leg is Gradually close to 0, thus controlling the rotation of the foot relative to the leg.

According to the control program of the second aspect of the present invention, the following function is provided to the computer mounted on the robot, that is, the function can alleviate the impact when the foot of the leg body touches the ground, and can also avoid The foot slips and spins so that it can walk or run stably, thus controlling the function of the robot.

In addition, the control program according to the second aspect of the present invention is characterized in that the following function is provided to the computer mounted on the robot as a function of controlling the movement of the legs. , in the state where the leg is still on the ground with the front end of the foot, the inclination angle of the foot relative to the ground measured by the foot inclination angle measurement function is toward the rear end of the foot that is farther away from the ground than the front end The positive side increases, thus controlling the function of the rotational movement of the foot relative to the leg.

In addition, the control program according to the second aspect of the present invention is characterized in that the following function is provided to the computer mounted on the robot as a function of controlling the movement of the legs. From the middle time of the period to the start time of the landing period, the inclination angle of the foot relative to the ground measured by the foot inclination angle measurement function gradually decreases to 0 from the angle at which the front end of the foot is farther away from the ground than the rear end. , thus controlling the function of the rotational movement of the foot relative to the leg.

In addition, the control program according to the second aspect of the present invention is characterized in that the following function is provided to the computer mounted on the robot as a function of controlling the movement of the legs. From the start time of the landing period to the start time of the landing period, the inclination angle of the foot relative to the ground measured by the foot inclination angle measurement function gradually increases toward the positive side and then gradually decreases, and then moves toward the front end of the foot than the rear. The negative side of the end away from the ground gradually increases, and then gradually decreases to 0, thus controlling the function of the rotation of the foot relative to the leg.

In addition, the control program according to the second aspect of the present invention is characterized in that a computer mounted on the robot is provided with a function of moving in the air phase with all the legs leaving the ground. , thus controlling the function of the movements of the legs of the robot.

Description of drawings

FIG. 1 is a schematic diagram showing the overall configuration of a bipedal mobile robot as a legged mobile robot according to an embodiment of the present invention.

FIG. 2 is a side view showing the configuration of the leg portion of each leg of the robot of FIG. 1 .

FIG. 3 is a block diagram showing the configuration of a control unit included in the robot of FIG. 1 .

FIG. 4 is a block diagram showing a functional configuration of a control unit in FIG. 3 .

FIG. 5 is an explanatory diagram illustrating a running gait of the robot of FIG. 1 .

FIG. 6 is an explanatory diagram of a foot position posture trajectory when the robot is running.

Fig. 7 is an explanatory diagram of the distance d between the sole of the foot and the ground.

Fig. 8 is an explanatory diagram of an angle θ formed between the sole of the foot and the ground.

FIG. 9 is a graph showing a setting example of a vertical component of a target ground reaction force.

FIG. 10 is a graph showing an example of setting a target ZMP.

Fig. 11 is a flowchart showing main program processing of the gait generating device included in the control unit in Fig. 3 .

Fig. 12 is a flowchart showing subroutine processing in the flowchart of Fig. 11 .

FIG. 13 is a graph showing an example of setting an allowable range of the ground reaction force horizontal component for a fixed gait.

Fig. 14 is a flowchart showing subroutine processing in the flowchart of Fig. 11 .

FIG. 15 is a graph showing an example of setting an allowable range of the horizontal component of the ground reaction force in the current gait.

FIG. 16 is a graph showing a setting example of the vertical component of the target ground reaction force in the walking gait.

FIG. 17 is a flowchart illustrating setting processing of a target ground reaction force vertical component in a walking gait.

Detailed ways

The embodiment of the legged mobile robot and its control program of the present invention will be described below with reference to the accompanying drawings.

A bipedal mobile robot (hereinafter referred to as robot) 1 shown in FIG. 1 has a pair of left and right leg bodies (leg links) 2 and 2 extending downward from an upper body 24 . The two leg bodies 2 and 2 are of the same structure and have 6 joints respectively. Its 6 joints are successively by from upper body 24 side: the joint 10R, 10L of crotch (waist) rotation (rotation) usefulness (relative to upper body 24 sideways swing direction use) (symbol R, L represent respectively and Meaning that the right leg body and the left leg body correspond; hereinafter the same), the joints 12R and 12L for rotation in the left and right (roll) direction (around the X axis) of the crotch (waist), the front and back of the crotch (waist) Joints 14R and 14L for rotation in the (pitch) direction (around the Y axis), joints 16R and 16L for rotation in the front-back direction of the knee, joints 18R and 18L for rotation in the front-back direction of the ankle, left and right of the ankle Joints 20R and 20L for rotation in the direction are formed.

At the bottom of the two joints 18R (L) and 20R (L) of the ankles of each leg body 2, a foot (foot) 22R (L) constituting the lower end of each leg body 2 is installed, 2. The uppermost part of 2 is equipped with the upper body (upper body) 24 through three joints 10R(L), 12R(L), and 14R(L) of the crotch of each leg body 2 . Inside the upper body 24, a control unit 26 and the like which will be described in detail later are accommodated. In addition, for ease of illustration, in FIG. 1 , the control unit 26 is drawn outside the upper body 24 .

The control unit 26 is composed of CPU, ROM, RAM, signal input circuit, signal output circuit, etc. as hardware, and the 'control program' of the present invention as software that provides the motion control function of the robot 1 to the hardware.

In each leg body 2 of said constitution, the hip joint (or waist joint) is constituted by joints 10R (L), 12R (L), 14R (L), the knee joint is constituted by joint 16R (L), and the foot joint (ankle joint) joint) is composed of joints 18R(L), 20R(L). In addition, the hip joint and the knee joint are connected by the thigh link 28R (L), and the knee joint and the foot joint are connected by the calf link 30R (L).

In addition, although not shown, a pair of left and right arm bodies are attached to both sides of the upper part of the upper body 24 , and a head is disposed on the upper end of the upper body 24 . Since these arms and heads are not directly related to the gist of the present invention, detailed descriptions are omitted, but the arms can be moved back and forth relative to the upper body 24 through a plurality of joints provided on each arm.

According to the above-mentioned structure of each leg body 2, the foot part (corresponding to the 'foot' in the present invention) 22R(L) of each leg body 2 is given 6 degrees of freedom with respect to the upper body 24 . Moreover, when the robot 1 moves, by driving the two legs 2 and 2 at an appropriate angle, they are combined into 6*2=12 (in this specification, this '*' is the Indicates the multiplication operation, and in the vector operation, it indicates the outer product) joints, which can make the two legs 22R, 22L perform desired motions. In this way, the robot 1 can arbitrarily move in three-dimensional space.

In addition, the position and speed of the upper body 24 that will be described later in this specification represent the predetermined position of the upper body 24 and its moving speed. The position of the central point between the hip joints, etc.). Similarly, the position and speed of each leg part 22R, 22L represent the position of the preset representative point of each leg part 22R, 22L, and its moving speed. In this case, in this embodiment, the representative point of each leg 22R, 22L is set, for example, on the bottom surface of each leg 22R, 22L (more specifically, it refers to the center of the ankle joint of each leg 2). point where a perpendicular line to the bottom surface of each leg portion 22R, 22L intersects the bottom surface, etc.).

As shown in FIG. 1 , a known six-axis force sensor 34 is attached below the ankle joints 18R(L) and 20R(L) of each leg body 2 and between the legs 22R(L). The 6-axis force sensor 34 is used to detect whether the foot 22R (L) of each leg body 2 is on the ground, and the ground reaction force (ground contact load) acting on each leg body 2, etc., and reacts the ground reaction force. Detection signals of the three-direction components Fx, Fy, and Fz of the translational force of the force and the three-direction components Mx, My, and Mz of the moment are output to the control unit 26 . In addition, the upper body 24 is provided with: an inclination sensor 36 for detecting the inclination (posture angle) of the upper body 24 relative to the Z axis (vertical direction (gravity direction)) and its angular velocity, and its detection signal is obtained from the inclination sensor 36 is output to the control unit 26 . In addition, although illustration of the detailed structure is omitted, each joint of the robot 1 is provided with: a motor 32 (refer to FIG. 3 ) for driving it, and a rotation amount of the motor 32 (rotation of each joint) Angle) encoder (rotary encoder) 33 (refer to FIG. 3 ), and a detection signal of the encoder 33 is output from the encoder 33 to the control unit 26 .

As shown in FIG. 2 , above each leg portion 22R (L), and between the six-axis force sensor 34, a spring mechanism 38 is provided, and on the sole (the bottom surface of each leg portion 22R (L)) A sole elastic body 40 made of rubber or the like is pasted. The soft mechanism 42 is constituted by these spring mechanisms 38 and the sole elastic body 40 . Specifically, the spring mechanism 38 consists of a square-shaped guide member (not shown) attached to the upper surface of the foot portion 22R (L) and an ankle joint 18R (L) (the ankle joint 20R is omitted in FIG. 2 ). (L)) and a six-axis force sensor 34 side and a piston-shaped member (not shown) that is housed in the guide member in a fine-movement manner via an elastic member (rubber or spring).

The leg part 22R (L) shown by the solid line in FIG. 2 shows the state when no ground reaction force is received. Once the legs 2 are subjected to the ground reaction force, the spring mechanism 38 and sole elastic body 40 of the flexible mechanism 42 are bent, and the foot 22R(L) is transferred to the position shown by the dotted line in the figure. The structure of the flexible mechanism 42 is not only for alleviating the impact of landing, but also plays an important role in improving the controllability of the robot 1 . In addition, since the details are already described in Japanese Unexamined Patent Application Publication No. 5-305584 previously filed by the present applicant, a detailed description thereof will be omitted.

In addition, although illustration is omitted in FIG. 1 , a joystick (joy-stick) (manipulator) 44 (see FIG. 3 ) for manipulating the robot 1 is provided outside the robot 1 as follows, namely , by operating the joystick 44 , a request for the gait of the robot 1 , such as rotating the robot 1 that is moving forward, is input to the control unit 26 as necessary. In this case, the requirements that can be input are: for example, the gait mode (walking, running, etc.) when the robot 1 moves, the landing position and posture or the landing time of the free leg, or the instruction data that stipulates these landing positions and postures or the landing time (For example, the moving direction and moving speed of the robot 1, etc.).

FIG. 3 is a block diagram showing the configuration of the control unit 26 . This control unit 26 is made up of microcomputer, and it has: the 1st computing device 60 and the 2nd computing device 62 that are constituted by CPU, A/D converter 50, counter 56, D/A converter 66, RAM54, ROM64, and carry out The bus 52 for data transmission and reception between these devices. In the control unit 26, the output signals of the 6-axis force sensor 34, the inclination sensor 36, the joystick 44, etc. of each leg body 2 are converted into digital values by the A/D converter 50, and then transmitted via the bus 52. Input to RAM54. In addition, the output of the encoder 33 (rotary encoder) of each joint of the robot 1 is input to the RAM 54 via the counter 56 .

The first computing unit 60 calculates the joint angle displacement command (command value of the displacement angle of each joint or the rotation angle of each motor 32) while generating the target gait as described later, and sends it to the RAM 54 . In addition, the second arithmetic unit 62 reads out the joint angle displacement command and the actual measurement value of the joint angle detected based on the output signal of the encoder 33 from the RAM 54, and calculates the operation amount necessary to drive each joint, And output to the motor 32 that drives each joint through the D/A converter 66 and the servo amplifier 32a.

FIG. 4 is a block diagram generally showing the functional configuration of the gait generation device and the control device of the robot 1 in this embodiment. Parts other than the "actual robot" part in FIG. 4 are constituted by processing functions executed by the control unit 26 (mainly the functions of the first computing device 60 and the second computing device 62). In addition, in the following description, when it is not necessary to distinguish the left and right of the leg body 2, the said symbol R, L is abbreviate|omitted.

As will be described below, the control unit 26 has a gait generating device 100 that freely and in real time generates and outputs a target gait described later. The gait generation device 100 is a gait generation device that constitutes each mechanism of the invention of the present application by its functions. The target gait output by the gait generation device 100 is composed of the target upper body position and posture track (the track of the target position and target posture of the upper body 24), the target foot position and posture track (the target position and target position of each foot 22). posture trajectory), target arm posture trajectory (orbit of each arm’s target posture), target total ground reaction center point (target ZMP) trajectory, and target total ground reaction force trajectory. In addition, when there are parts that can move relative to the upper body 24 other than the legs 2 and arms, the target position posture trajectory of the movable parts can be added to the target gait.

Here, the "trajectory" of the gait is a pattern (time-series pattern) showing a temporal change, and in the following description, it may be referred to as a "pattern" instead of the "trajectory". In addition, 'posture' indicates a spatial orientation. Specifically, for example, the upper body posture is determined by the inclination angle (posture angle) of the upper body 24 in the left-right direction (around the X-axis) relative to the Z-axis (vertical axis) and the inclination angle (posture angle) of the upper body 24 in the front-back direction (around the Y-axis). The foot posture is represented by a biaxial spatial azimuth fixedly set on each foot 22 . In this specification, an upper body posture may also be referred to as an upper body posture angle.

In addition, in the following description, 'target' is often omitted when there is no possibility of misunderstanding. In addition, the gait related to the movement of the robot 1 such as the foot position and posture, the upper body position and posture, and other components other than the ground reaction force-related components in the gait is collectively referred to as "movement". In addition, the ground reaction force (ground reaction force composed of translational force and moment) acting on each leg 22 is referred to as "each leg ground reaction force", and all (two) legs 22R, 22L of the robot 1 are The resultant force of the 'ground reaction force of each foot' is called the 'total ground reaction force'. However, in the following description, since the ground reaction force of each foot is hardly mentioned, unless otherwise specified in advance, 'ground reaction force' and 'total ground reaction force' are treated in the same meaning.

The target ground reaction force is generally represented by the point of action and the translational force and moment acting on the point. Since the point of action can be anywhere, even the same target ground reaction force can have countless representations. position) as the point of action to represent the target ground reaction force, the moment component of the target ground reaction force is zero except for the vertical component (moment around the vertical axis (Z axis)). In other words, the horizontal component of the moment of the target ground reaction force around the center point of the target ground reaction force (moment around the horizontal axis (X-axis and Y-axis)) is zero.

In addition, in the gait that satisfies the dynamic balance condition, since the ZMP (calculated based on the target trajectory of the robot 1 is the point where the following moment is zero except for the vertical component, the moment refers to the point calculated based on the target trajectory The resultant force of inertial force and gravity acting around its point) is consistent with the center point of the target ground reaction force, so replacing the orbit of the center point of the target ground reaction force to provide the target ZMP orbit can be said to be the same (see details in this case Special Request No. 2000-352011 issued by the applicant, etc.).

Against this background, the target gait is defined as follows in the specification of Japanese Patent Application No. 2000-352011.

a) The generalized target gait refers to a set of target motion trajectories and target ground reaction force trajectories during one or more steps.

b) Target gait in a narrow sense refers to a set of target motion trajectory and its ZMP trajectory during 1 step.

c) A series of gaits refers to the situation in which several gaits are connected.

During walking, if the vertical position (upper body height) of the upper body 24 of the robot 1 is determined according to the method for determining the height of the upper body proposed by the applicant in JP 10-86080, then the translational ground reaction The vertical component of the force is then determined. In addition, the horizontal position track of the upper body of the robot 1 is determined in such a way that the horizontal component of the torque generated by the resultant force of inertial force and gravity generated by the movement of the target gait around the target ZMP is zero, thereby translating the ground reaction force Level components are also determined. Therefore, in the specification of Japanese Patent Application No. 2000-352011, it is sufficient to have the target ZMP as a physical quantity that should be clearly set regarding the ground reaction force of the target gait. Therefore, as the definition of the target gait in a narrow sense, the above b) is sufficient. On the other hand, in the running gait of the robot 1 described in this embodiment (details will be described later), the vertical component of the ground reaction force (vertical component of the translational ground reaction force) is also important in terms of control. Therefore, in the present invention, after clearly setting the target trajectory of the vertical component of the ground reaction force, the trajectory such as the target upper body vertical position of the robot 1 is determined. Therefore, in this specification, the following b') is used as the definition of the target gait in a narrow sense.

b') The target gait in the narrow sense refers to a set of target motion trajectory during one step, its target ZMP trajectory, and the vertical component trajectory of the target translational ground reaction force.

In this specification, hereinafter, for easy understanding, unless otherwise specified, the target gait uses the meaning of the target gait in the narrow sense of b') above. In this case, the "one step" of the target gait refers to the time when one leg 2 of the robot 1 touches the ground until the other leg 2 touches the ground. In addition, in the following description, "the vertical component of the ground reaction force" means "the vertical component of the translational ground reaction force", and the vertical component (the component around the vertical axis) of the moment in the ground reaction force uses " The term "moment" is used to distinguish it from the "vertical component of ground reaction force". Likewise, 'horizontal component of ground reaction force' means 'horizontal component of translational ground reaction force'.

In addition, of course, the two-leg support period in the gait refers to the period when the robot 1 supports its own weight with two legs 2 and 2, and the single-leg support period refers to the period when only any one leg 2 supports the robot 1’s own weight. During the period in the air, the period in the air refers to the period during which the two leg bodies 2, 2 have left the ground (floating in the air). In the single-leg support period, the side leg body 2 that does not support the self-weight of the robot 1 is called a 'free leg'. In addition, in the running gait described in this embodiment, there is no two-leg stance phase, but a single-leg stance phase (landing phase) and an air phase are repeatedly and alternately performed. In this case, in the air phase, both legs 2, 2 do not support the weight of the robot 1, but in the single leg support phase before the air phase, even if the leg body 2 as the free leg and the leg body 2 as the supporting leg are in the air phase. The aerial period is also called the free leg and the supporting leg respectively.

Taking the running gait shown in FIG. 5 as an example, the outline of the target gait generated by the gait generating device 100 will be described. In addition, other definitions and details about gait have been previously applied for by the applicant of the present case and recorded in JP-A-10-86081, so the following content that is not described in JP-A-10-86081 Explain for the master.

First, the running gait shown in FIG. 5 will be described. This running gait is the same as a normal running gait of a person. In this running gait, the single-leg support period and the aerial period in which both legs 2 and 2 are floating in the air are repeated and alternately performed, wherein the single-leg support period refers to only one of the left and right legs of the robot 1 The foot 22 of the leg body 2 (supporting leg) is in the supporting phase of landing (touching the ground).

This running gait is a gait that changes in the following chronological order, that is, as shown in FIG. The middle moment (t=t1) of the single-leg support period when the foot 22L is almost on the ground with its entire bottom surface; as shown in Figure 5(b), the right leg 2R in front of the left leg 2L is Move to the front, and the end time of the single-leg support period (~the start time of the next mid-air period) (t=t2 ); as shown in Figure 5 (c), the right leg body 2R is in the front of the left leg body 2L, and the two leg bodies are in the air period (t=t3) leaving the ground; as shown in Figure 5 (d), the right leg The foot 22L of body 2R is almost the start moment of the single-leg support period (~the end moment of the previous sky period) (t=t4) when the entire surface of its bottom surface comes to the ground; as shown in Figure 5 (e), the position The left leg body 2L in front of the right leg body 2R is moved to the front, and the foot portion 22R of the left leg body 2R is on the ground and moves obliquely in the mode of its rear end side rising again (～ The start moment of the next air period) (t=t5); as shown in Figure 5 (f), the left leg body 2L is in front of the right leg body 2R, and the two leg bodies are in the air period (t=t6) leaving the ground and as shown in Fig. 5 (g), the starting moment of the single-leg support period (~the end moment of the previous sky period) (t=t7 ).

Taking the running gait in FIG. 5 into consideration, the basic outline of the target gait generated by the gait generating device 100 will be described. As will be described later in detail, when the gait generation device 100 generates the target gait, the grounding position and posture (planned landing position and posture) and the landing time (planned landing time) of the foot 22 on the free leg side are used to generate the target gait. The basic request value (request parameter) is supplied to the gait generation device 100 in accordance with the required operation of the joystick 44 and the like. Also, the gait generation device 100 generates a target gait using its required parameters. To be more specific, the gait generation device 100 determines parameters (called gait parameters) according to the required parameters, and the gait parameters are the target foot position posture track, the target ground, and the target ground for specifying the target gait. The reaction force is a part of the target gait such as the vertical component trajectory, and then the gait parameters are used to sequentially determine the instantaneous value of the target gait, thereby generating a temporal pattern of the target gait.

In this case, the target foot position posture trajectory (more specifically, the target trajectory of each component (X-axis component, etc.) in the space of the position and posture of the foot) is used, for example, in Patent No. The applicant's proposed finite-time tuning filter is generated for each foot 22 . The finite-time tuning filter is a filter whose transfer function is represented by a first-order delay filter with a variable time constant, that is, in the form of 1/(1+τs) (τ is a variable time constant. Hereinafter, the filter is referred to as Unit filter) is a filter in which multiple stages (3 or more stages in this embodiment) are connected in series, and it can generate and output a trajectory that reaches a specified value at a desired specified time. In this case, the time constant τ of the unit filter of each stage is sequentially and variablely set according to the remaining time until the above-mentioned specified time after the output generation of the finite-time tuning filter is started. More specifically, τ is set in the following manner, that is, with the shortening of the remaining time, the value of τ decreases from a predetermined initial value (>0), and finally at a specified time when the remaining time is 0, τ value becomes 0. And, a step input of a height corresponding to the specified value (more specifically, an amount of change from the initial value of the output of the finite-time tuning filter toward the specified value) is supplied to the finite-time tuning filter. Such a finite-time tuning filter not only generates an output that reaches a specified value at a specified time, but also makes the change speed of the output of the finite-time tuning filter at a specified time zero or almost zero. Especially when three or more stages (or three stages) of unit filters are connected, the change acceleration (differential value of the change speed) of the output of the finite time tuning filter can be made zero or almost zero.

The generation of the foot position and posture trajectory (the position and posture trajectory from the time the foot 22 hits the ground to the next landing) using such a finite time tuning filter can be performed, for example, as follows. For example, the target foot position trajectory in the X-axis direction (front-rear direction) is generated as follows. That is, according to the position in the X-axis direction of the next scheduled landing position of each foot 22 determined by the required parameters (more specifically, the position of one of the next scheduled landing positions relative to the X-axis direction of the previous landing position) variation (move amount). It is equivalent to the specified value), to determine the height of the step input supplied to the finite-time tuning filter, and after the time constant τ is initialized to a prescribed initial value, the determined The step input is supplied to the finite-time tuning filter, and generation of the trajectory of the position in the X-axis direction of the leg portion 22 is started. Then, at the time of generation of this trajectory, the time constant τ is sequentially and variably set so as to decrease from the initial value to 0 until the scheduled landing time of the foot 22 (which corresponds to the designated time). Accordingly, a trajectory of the position of the foot portion 22 in the X-axis direction such as reaching the scheduled landing position at the scheduled landing time is generated.

In addition, the target foot position trajectory in the Z-axis direction (vertical direction) is generated as follows, for example. That is, first, according to the next scheduled landing position and scheduled landing time of the foot 22, the position in the Z-axis direction of the foot 22 when the height (vertical position) of the foot 22 is the maximum (hereinafter referred to as highest point position), and the time of arrival at its highest point position. Also, the height of the step input supplied to the finite-time tuning filter is determined according to its highest point position (which corresponds to the specified value), and after the time constant τ is initialized, its determined step input is supplied to The filter is tuned in a finite time, and the foot position trajectory in the Z-axis direction to the highest point position is sequentially generated. At this time, the time constant τ is sequentially and variably set so as to decrease from the initial value to 0 before the arrival time (corresponding to the designated time) at which the highest point position is reached. In addition, when the generation of the trajectory of the position in the Z-axis direction that reaches the highest point position is completed, the time constant τ is initialized, and the step input so far and the step input of reverse polarity (more specifically, the same as that from The step input of the reverse polarity of the height corresponding to the change in the Z-axis direction (which is equivalent to the specified value) from the highest point position to the next scheduled landing position) is input to the finite time tuning filter to sequentially generate The trajectory of the foot position in the Z-axis direction from the highest point position to the predetermined landing position. At this time, the time constant τ is sequentially and variably set so as to decrease from the initial value to 0 before reaching the scheduled landing time of the foot portion 22 .

Accordingly, the distance d between the sole of the left leg body 2L and the ground in the running gait of FIG. 5 varies as shown in FIGS. location track.

In addition, in generating the trajectory of the foot position in the Z-axis direction, the time constant τ can be variably set so as to continuously decrease from the initial value to 0 from the trajectory generation start time to the scheduled landing time of the foot 22, And at or near the time of arrival at the highest point position, by switching the polarity of the step input into reverse polarity, the foot position trajectory in the Z-axis direction is generated. In this case, although the leg portion 22 cannot reach the desired highest point position with high precision, there is no problem in reaching the predetermined landing position at the predetermined landing time.

Even the foot posture track can be generated using a finite time tuning filter in the same way as the above-mentioned foot position track. In this case, among the spatial components of the foot posture, the posture angle change is monotonous (monotonically increasing or monotonically decreasing), as long as it can be combined with the above-mentioned generation of the foot position trajectory in the X-axis direction. In the same way, the foot pose track can be generated. In addition, with regard to components such as a maximum value or a minimum value in a posture angle change, it is only necessary that the foot posture trajectory can be generated in the same way as the above-mentioned generation of the foot position trajectory in the Z-axis direction.

Accordingly, the posture of the left foot 22L in the running gait of FIG. 5 changes as shown in FIGS. 6 and 8 , thereby generating a foot posture trajectory in the Z-axis direction. Specifically, from the inclined state in which the front end (toe) of the sole of the foot during the lift-off period is higher than the rear end (heel) with respect to the ground, the front end is continuously tilted and moved while the front end is lowered. The inclination state of the leg portion 22L is changed so as to be substantially parallel to the ground before transitioning from this ground-off period to the next landing period. The angle θ is defined as negative (-) when the sole of the foot is tilted forward relative to the ground, and positive (+) when the sole is lowered forward.

The robot 1 of the present invention controls the inclination state of the legs 22 in the air (see FIG. 5( c ), ( f )) so that the angle θ between the soles of the feet and the ground becomes zero. That is, as shown in FIG. 5( g ) and FIG. 6 , the angle θ of the left foot portion 22L is controlled to be zero at the latest at time t=t7 when the air phase transitions to the landing phase.

Alternatively, the angle θ may be defined as a function θ(d) of the distance d. In the following occasions, as mentioned above, the inclination state of the foot portion 22 is controlled in such a manner that it is approximately parallel to the ground before changing to the next landing stage while the sole of the foot is continuously tilted during the lift-off period (refer to In the case of FIG. 6 ), θ(d=0)=0, and δθ/δd>0 before the transition from the ground lift phase to the ground phase. Alternatively, the angle θ may be controlled to be 0 during the air phase, and the angle θ may be maintained at 0 without any change until the transition time to the landing phase.

In addition, the target foot position posture trajectory generated by the finite time tuning filter as described above is the target position posture trajectory of each foot 22 in the support leg coordinate system fixed to the ground which will be described later.

The target foot position posture trajectory generated as described above is such that the position of each foot 22 starts to move while gradually accelerating from its initial ground contact state (the state at the initial moment of the target gait) toward the predetermined ground position. to generate. In addition, the target foot position posture trajectory is generated in such a way that the rate of change of the position gradually decreases to 0 or almost 0 until the scheduled landing time, and then stops at the scheduled landing position at the scheduled landing time. . Accordingly, the ground speed of each foot 22 at the moment of landing (the speed of change of the position of each foot 22 in the support leg coordinate system fixed to the ground) is zero or almost zero. Therefore, in the running gait, even if all the leg bodies 2, 2 touch the ground simultaneously from the state of being in the air (the state of being in the air), the landing impact will be small.

In the running gait, according to the gravity acting on the robot 1, the vertical speed of the upper body 24 becomes downward from the later stage of the aerial phase, and it is still in a downward state when it lands. Therefore, as described above, the target foot position posture trajectory is generated so that the ground speed of each foot 22 at the moment of landing is 0 or almost 0, and the target position of the upper body 24 is generated so as to satisfy the dynamic balance condition described later. Position posture track, at this time, just before landing, the relative speed of the foot 22 on the free leg side with respect to the upper body 24 becomes upward. That is, at the moment of landing in the running gait, the target gait of the robot 1 is a gait in which the leg body 22 on the side of the free leg is retracted to the upper body 24 side while landing. In other words, in the target gait of this embodiment, the moment the robot 1 hits the ground, the ground speed of the foot 22 on the side of the free leg is 0 or almost 0. The portion 22 is lifted to land. According to this, the ground impact becomes small, thereby preventing the ground impact from becoming excessively large.

In addition, in this embodiment, since the finite-time tuning filter is formed by connecting unit filters in series of more than three stages (for example, three stages), not only each leg portion 22 The speed of the foot (the change speed of the foot position) is 0 or almost 0, and the acceleration of each foot 22 is also 0 or almost 0 at the predetermined landing time and stops. That is, the ground acceleration at the moment of landing is also zero or almost zero. Therefore, the ground impact becomes smaller and smaller. Especially when the actual moment of landing of the robot 1 deviates from the target moment of landing, the impact will not increase much. As a supplement, regarding the aspect of making the ground speed of each leg portion 22 be 0 or almost 0 at the predetermined moment of landing, the number of stages of the unit filter of the finite time tuning filter can also be 2 stages, but in this case, in The acceleration of each foot 22 at the predetermined moment of landing is generally not zero.

In addition, the posture of the legs is temporarily maintained constant after each leg 22 lands on the ground with almost the entire bottom surface thereof at the scheduled landing time. Thus, the time when the foot 22 touches the ground with almost the entire bottom surface thereof is set as the specified time, and the foot posture trajectory is generated by the finite time tuning filter.

In addition, in the present embodiment, although the finite time tuning filter is used to generate the foot position trajectory, it is also possible to generate the target foot position trajectory using a function such as a polynomial set in the following manner. The change speed of the foot position at the predetermined landing time is 0 or almost 0 (the time differential value of the foot position is 0), and then the change acceleration (time differential value of the change speed) of the foot position at the predetermined landing time value) is set to be 0 or almost 0. The same applies to the generation of the target leg posture track. However, when generating the target foot posture trajectory, as described above, when almost the entire bottom surface of each foot 22 is placed on the ground, the change speed of the posture of each foot 22 and the corresponding Functions such as polynomials are set in such a way that the variable acceleration is zero or almost zero.

The target ground reaction force vertical component trajectory is set as shown in FIG. 9 , for example. In this embodiment, the shape of the track of the vertical component of the target ground reaction force in the running gait (specifically, the shape during the single-leg support phase) is determined as a trapezoidal shape (the increase in the vertical component of the ground reaction force The shape of the side bulge), the height of the trapezoid and the moment of the inflection point are used as the gait parameters specifying the vertical component trajectory of the target ground reaction force to determine these gait parameters (parameters of the vertical component trajectory of the ground reaction force). In addition, the vertical component of the target ground reaction force is fixedly set to zero during the air phase of the running gait. As described in this example, it is preferable to set the trajectory of the vertical component of the target ground reaction force to be substantially continuous (so that the value does not become discontinuous). This is for smooth motion of the joints of the robot 1 when controlling the ground reaction force. In addition, the term "substantially continuous" means that when a continuous track (a true continuous track) is displayed analogously and digitally in a discrete-time system, the jump in values that inevitably occurs does not cause the track to lose continuity.

The target ZMP orbit is set as follows. In the running gait of Fig. 5, as previously mentioned, the supporting leg side foot 22 is almost on the ground with the entire surface of its bottom surface, and then kicks out and flies into the air with the tiptoe of its supporting leg side foot 22, Finally, almost land on the ground with the entire surface of the bottom surface of the free leg side foot portion 22 . Therefore, as shown in the upper section figure among Fig. 10, the target ZMP orbit in the single-leg support period is set as follows, that is, with the middle position of the heel and toe of the supporting leg side foot 22 as the initial position, then, The period during which almost the entire bottom surface of the supporting leg side foot part 22 touches the ground is maintained constant, and thereafter, the toe of the supporting leg side foot part 22 moves toward the time of leaving the ground. Here, the upper diagram in FIG. 10 shows the target ZMP trajectory in the X-axis direction (front-back direction), and the lower diagram in FIG. 10 shows the target ZMP trajectory in the Y-axis direction (left-right direction). In addition, the target ZMP trajectory in the Y-axis direction in the single-leg stance phase is set at the same position as the center position of the ankle joint of the supporting leg 2 in the Y-axis direction, as shown in the lower diagram of FIG. 10 . .

In the running gait, and then after the end of the single-leg support period, the two-leg bodies 2, 2 leave the ground, and the vertical component of the ground reaction force is 0. When the vertical component of the ground reaction force is 0, that is, in the air phase, the overall center of gravity of robot 1 is in free fall, and the change of angular momentum around the overall center of gravity is zero. At this time, since the moment of the resultant force of the gravity and the inertial force acting on the robot 1 is 0 at any point on the ground, the target ZMP is uncertain. In other words, any point on the ground satisfies the ZMP condition of the so-called "action point at which the horizontal component of the moment of the resultant force of gravity and inertial force is 0". In other words, setting the target ZMP at any point can satisfy the so-called dynamic balance condition that the horizontal component of the moment of the resultant force acting around the target ZMP is 0. Therefore, the target ZMP may be set discontinuously. For example, in the air phase, the target ZMP can be set in such a way that it does not move from the target ZMP position when it is off the ground (at the end of the single-leg support phase), and at the end of the air phase, it can also be set to the target ZMP position when it is on the ground. The target ZMP trajectory is set in a manner of discontinuous (stepwise) movement. However, in the present embodiment, as shown in the upper diagram in FIG. 10 , the position in the X-axis direction of the target ZMP orbit in mid-air is changed as follows: The toe of the side foot portion 22 moves continuously to the middle position between the heel and the toe of the free leg side foot portion 22 . In addition, as shown in the lower diagram in FIG. 10 , the Y-axis direction position of the target ZMP trajectory in the air phase changes as follows, that is, until the next free leg side leg body 2 touches the ground, the distance from the ankle of the supporting leg side leg body 2 The Y-axis direction position of the joint center moves continuously to the Y-axis direction position of the ankle joint center of the free leg side leg body 2 . That is, the target ZMP trajectory is made continuous (substantially continuous) throughout the duration of the gait. And, as will be described later, in order to make the moment of the resultant force of the gravity and inertial force around the target ZMP (except for the vertical component) zero to generate the target gait (more specifically, it refers to adjusting the target upper body position posture trajectory ).

In addition, in this embodiment, the position and time of the inflection point of the target ZMP trajectory as shown in FIG. 10 are set as ZMP trajectory parameters (parameters defining the target ZMP trajectory). In addition, the meaning of "substantially continuous" in the above-mentioned ZMP orbit is the same as that in the case of the above-mentioned ground reaction force vertical component orbit.

The ZMP orbital parameters are determined in such a way that they are sufficiently stable and do not change drastically. Here, the state where the target ZMP exists near the center of the smallest convex polygon (so-called support polygon) including the contact surface of the robot 1 is called sufficiently stable (see JP-A-10-86081 for details). The target ZMP orbit in FIG. 10 is an orbit set so as to satisfy this condition.

In addition, the target arm posture is represented by a relative posture with respect to the upper body 24 .

In addition, the target upper body position and posture, the target foot position and posture, and the reference upper body posture described later are expressed in a world (global) coordinate system. The world coordinate system refers to the coordinate system fixed on the ground as mentioned above. As the world coordinate system, more specifically, a support leg coordinate system described later is used.

The gait generation device 100 in this embodiment is a target gait (target gait in the narrow sense) of one step from the time when one leg 2 of the robot 1 lands on the ground to the other leg 2 lands on the ground. The target gait of one step of the robot 1 is sequentially generated in units of . Therefore, in the running gait of FIG. 5 generated in this embodiment, the target gait is sequentially generated from the beginning of the single-leg support phase to the end of the immediately following aerial phase (the next single-leg phase). The target gait at the beginning of the leg stance phase). Here, the newly generated target gait is called "this gait", the next target gait is called "next gait", and the next target gait is called "next step". secondary gait'. In addition, the target gait generated before the "this gait" is referred to as the "previous gait".

In addition, when the gait generation device 100 newly generates the gait this time, the expected landing position and posture of the free leg side foot 22 two steps ahead of the robot 1, and the required value (requirement) at the scheduled landing time are used as relative gait values (requirements). The required parameters of the posture are input to the gait generation device 100 (or the gait generation device 100 reads the required parameters from the storage device). Then, the gait generating device 100 generates target upper body position and posture trajectories, target foot position and posture trajectories, target ZMP trajectories, target ground reaction force vertical component trajectories, target arm posture trajectories, etc. using these required parameters. At this time, to ensure the continuity of walking, a part of the gait parameters specifying these trajectories was appropriately modified.

Taking the running gait in FIG. 5 as an example, the gait generation process of the gait generation device 100 will be described in detail below with reference to FIGS. 11 to 15 . FIG. 11 is a flowchart (structured flowchart) showing gait generation processing executed by the gait generation device 100 .

First, in S010, various initialization operations such as initializing the time t to 0 are performed. This processing is performed when the gait generation device 100 is activated, for example. Next, proceed to S014 via S012, and the gait generating device 100 waits for the time insertion of each control cycle (operation processing cycle in the flow chart of FIG. 11 ). The control period is Δt.

Then, enter S016, judge whether it is the transition point of gait, if it is the transition point of gait, enter S018, and when it is not the transition point, enter S030. Here, the above-mentioned "gait transition point" indicates the timing at which the generation of the previous gait is completed and the generation of the current gait is started. For example, the next control cycle of the control cycle after the generation of the previous gait is the gait the conversion point.

When entering S018, the time t is initialized to 0, and then enters S020 to read the next gait support leg coordinate system, the next gait support leg coordinate system, the current gait cycle and the next gait cycle. These supporting leg coordinate systems and gait cycle are determined according to the above-mentioned required parameters. That is, in the present embodiment, the requested parameters supplied to the gait generating device 100 from the joystick 44 or the like include: the predetermined landing position and posture of the free leg side foot 22 up to 2 steps forward (the position and posture in the state of turning without slipping). Foot position and posture, so that after the foot 22 hits the ground, the sole of the foot is almost in contact with the ground with the entire surface), and the required value at the time of scheduled landing, the required value of the first step and the required value of the second step are respectively corresponding The values of the current gait and the next gait are supplied to the gait generation device 100 before the start of generation of the current gait (the transition point of the above-mentioned S016 gait). In addition, these required values may be changed during the generation of this gait.

And, corresponding to the request value of the predetermined landing position posture of the free leg side foot 22 (free leg side foot 22 under this gait) of the first step in the above-mentioned request parameters, determine the next gait supporting leg Coordinate System.

Afterwards, enter S022, the gait parameters of the fixed rotation gait are determined by the gait generating device 100, and the fixed rotation gait is used as a hypothetical periodic gait following this gait. The gait parameters include: foot trajectory parameters specifying a target foot position posture trajectory in a fixed rotation gait, reference upper body posture trajectory parameters defining a reference upper body posture trajectory, and arm trajectory parameters specifying a target arm posture trajectory , specify the ZMP orbit parameters of the target ZMP orbit, and specify the ground reaction force vertical component orbit parameters of the target ground reaction force vertical component orbit. In addition, the gait parameters also include parameters that specify the allowable range of the horizontal component of the target ground reaction force.

In addition, in this specification, "fixed rotation gait" is used as the meaning of periodic gait, and this periodic gait means: when performing its gait repeatedly, at the critical point of gait (in this embodiment, The critical state of each gait), the motion state of robot 1 (states such as foot position and posture, upper body position and posture) will not produce discontinuity. In the future, "fixed rotation gait" will sometimes be referred to simply as "fixed gait".

The fixed rotation gait is temporarily created for the purpose of specifying the divergent component, upper body vertical position velocity, upper body posture angle, and angular velocity at the terminal time of this gait by the gait generation device 100 This fixed rotation gait is not the gait output from the gait generation device 100 as it is, depending on the motion state of the robot 1 .

In addition, "divergence" means that the position of the upper body 24 of the biped mobile robot 1 deviates to a position away from the positions of the two legs 22 , 22 . The value of the so-called divergence component means that the position of the upper body 24 of the bipedal mobile robot 1 is far from the positions of the two feet 22, 22 (more specifically, it is the world coordinate system set on the contact surface of the supporting leg side foot 22). (the origin of the supporting leg coordinate system)) the value in the case.

Returning to the original topic, in S022, the following processing is performed according to the flowchart shown in FIG. 12 .

First, at S100, the foot trajectory in the gait parameters of the fixed gait is determined in such a way that the foot position posture trajectory is arranged in the order of the current gait, the first rotation gait, and the second rotation gait parameter. The specific setting method is explained below. In addition, in the following description, the leg part 22 of the leg body 2 of a supporting leg side is called a supporting leg part, and the leg part 2 of the leg body 2 of a free leg side is called a free leg part. In addition, the 'initial' and 'terminal' of the gait represent the beginning moment, the end moment of the gait, or the instantaneous gait at these moments, respectively.

The foot trajectory parameters are composed of the respective positions and postures of the supporting leg and the free leg, and the gait cycle of each rotational gait. At this time, the supporting leg and the free leg refer to the first rotational gait and the initial and terminal respective support leg and free leg of the second rotational gait. The initial free leg and foot position and posture of the first rotational gait among the foot track parameters refer to the position and posture of the terminal supporting leg and foot of this gait viewed from the supporting leg coordinate system of the next gait. In this case, in the running gait, the supporting leg portion 22 at the end of this gait is moving in the air. Moreover, the position and posture of the supporting leg foot at the terminal of this gait is obtained by using the finite-time tuning filter to generate a trajectory of the foot position and posture from the bottom up until the terminal of this gait. Secondary gait support leg coordinate system), and the foot position posture orbit refers to: from the initial support leg foot position posture of this gait (=previous gait terminal free leg foot position posture) to the next gait The position and posture of the terminal free leg and foot, and the position and posture of the terminal free leg and foot of the next gait are determined according to the following requirement value or according to the next gait support leg coordinate system corresponding to the requirement value, the requirement Value refers to the requirement value of the predetermined landing position posture of the free leg side foot portion 22 of the 2nd step among the described request parameters (the predetermined landing position posture in the next gait of the supporting leg portion 22 of this gait) Required value).

Next, enter S102, determine the reference upper body posture trajectory parameters, and the reference upper body posture trajectory parameters are used to specify the reference upper body posture trajectory that the target upper body posture should follow. Although as long as the initial stage of the fixed gait (the initial stage of the first rotational gait) and the terminal (the terminal of the second rotational gait) are connected to the ground (so that the initial stage of the fixed gait, the posture angle and angular velocity of the reference upper body posture of the terminal It is not necessary to be a certain posture if the reference upper body posture is set in accordance with the above, but, in this embodiment, for easy understanding, the reference upper body posture is set as an upright posture (vertical posture). That is, in the present embodiment, the reference upper body posture is set as the upright posture throughout the fixed gait. Therefore, in the present embodiment, the angular velocity and angular acceleration of the posture angle of the reference upper body posture are zero.

Thereafter, enter S104, determine the arm posture trajectory parameters, more specifically, determine the arm posture trajectory parameters except the angular momentum changes of the two arms around the vertical axis (or upper body trunk axis). For example, determine the arm posture trajectory parameters such as the relative height of the fingertips of the arm body relative to the upper body 24 and the relative center of gravity position of the whole arm. In addition, in the present embodiment, the relative center of gravity position of the entire arm is set to be kept constant with respect to the upper body.

Afterwards, enter S106, and set the orbital parameters of the vertical component of the ground reaction force. In this case, the trajectory of the vertical component of the ground reaction force specified by this parameter becomes substantially continuous (value There will be no stepped jump) track, so set the ground reaction force vertical component track parameters. In this mode, in any gait of the first rotation gait and the second rotation gait, the vertical component of the ground reaction force becomes trapezoidal in the single-leg support phase, and the vertical component of the ground reaction force becomes trapezoidal in the air phase. Straight components are maintained at 0. And, the timing of the inflection point of this pattern and the height (peak value) of the trapezoidal portion are set as the ground reaction force vertical component orbit parameters.

Then, enter S108, set the allowable range [Fxmin, Fxmax] of the horizontal component of the ground reaction force as shown in Figure 13 according to the vertical component trajectory of the ground reaction force set as above (further specifically specify parameters). The broken line on the negative side in FIG. 13 represents the allowable lower limit value Fxmin of the horizontal component of the ground reaction force, and the broken line on the positive side represents the allowable upper limit value Fxmax of the horizontal component of the ground reaction force. These setting methods are supplemented below. Next, a case where the ground is horizontal will be described.

Although the horizontal component of the ground reaction force is generated by the friction between the ground and the feet 22, the friction is not generated indefinitely, but has a limit. Therefore, in order for the actual robot 1 to move in accordance with the generated target gait without slipping, the level component of the ground reaction force of the target gait must always be within the friction limit. Therefore, in order to satisfy this condition, the allowable range of the horizontal component of the ground reaction force is set, as described later, so that the horizontal component of the ground reaction force of the target gait is within the allowable range, and the target gait is generated in this way.

If the coefficient of friction between the ground and the leg 22 is μ, Fxmin must always be set above −μ*the vertical component of the ground reaction force, and Fxmax must be set below μ*the vertical component of the ground reaction force. The simplest setting method is set by the following formula. Here, ka is a positive constant smaller than 1.

Fxmin＝-ka*μ*vertical component of ground reaction force

Fxmax＝ka*μ*vertical component of ground reaction force ... Formula 12

The permissible range of the horizontal component of the ground reaction force in FIG. 13 is an example set according to Equation 12. As parameters specifying the allowable range of the horizontal component of the ground reaction force, although the value and time at the inflection point such as the trapezoidal waveform in Fig. Only the value of (ka*μ) in Equation 12 may be set as a parameter.

Then, go to S110, and set the ZMP trajectory parameter, which is the ZMP trajectory for specifying the fixed gait after the combination of the first rotation gait and the second rotation gait. In this case, the target ZMP orbit is set in such a manner that it is sufficiently stable and does not change rapidly as described above.

To be more specific, in the running gait of FIG. 5 , after almost touching the ground with the entire bottom surface of the supporting leg portion 22 , the supporting leg portion 22 is maintained to touch the ground with almost the entire bottom surface thereof. , thereafter, only the toe of the supporting leg foot 22 touches the ground. Then, then kick out with the tiptoes of the supporting legs and feet 22, fly into the air, and almost land on the ground with the entire surface of the bottom surface of the free legs and feet 22 at last. Additionally, the target ZMP must be within the touchdown. Therefore, in this embodiment, as shown in the upper diagram of FIG. 10 , the positions in the X-axis direction of the respective target ZMPs of the first rotational gait and the second rotational gait of the fixed gait are as follows Setting, that is, with the intermediate position between the heel and the toe of the supporting leg foot 22 as the initial position, after being temporarily maintained at a certain level, the foot 22 moves toward the toe before the toe touches the ground, and thereafter, until the time of leaving the ground , stop at the tiptoe of the supporting leg foot 22. After that, the target ZMP is set as follows, that is, as previously mentioned, until the next free leg 22 hits the ground, the target ZMP moves continuously from the toe of the supporting leg 22 to the heel and toe of the free leg 22 middle position. And, the time and position of the inflection point of the target ZMP trajectory are set as ZMP trajectory parameters. In this case, the timing of the inflection point is set according to the gait cycle of the first turning gait and the second turning gait, and the gait periods of the first turning gait and the second turning gait are set according to The required parameters are determined; the position of the inflection point is based on the position and posture of the next gait supporting leg coordinate system and the next gait supporting leg coordinate system, or according to the first step and the first step of specifying the required parameters of these coordinate systems In the second step, the free leg side foot is set according to the required value of the position and posture of the free leg side. In addition, the position in the Y-axis direction of the ZMP track is set to be the same as that shown in the lower diagram of FIG. 10 . More specifically, the trajectory of the Y-axis position of the target ZMP in the first rotational gait is set in the same pattern as the lower diagram in FIG. 10 , and the Y-axis position of the target ZMP in the second rotational gait The trajectory of is a trajectory having the same shape as that of the first rotation gait, and is set so as to be connected to the terminal of the trajectory.

After performing the processing shown in S010 to S022 in FIG. 11 , the process proceeds to S024 to calculate the initial state of the fixed gait. The initial states calculated here are: initial upper body horizontal position velocity of fixed gait (initial upper body position and initial upper body velocity in the horizontal direction), initial upper body vertical position velocity (initial upper body velocity in the vertical direction Position and initial upper body velocity), initial divergent component, initial upper body posture angle and its angular velocity. The computation of this initial state can be done heuristically.

Next, enter S026 in FIG. 11 , determine (partially determine temporarily) the gait parameters of this gait. In S026, more specifically, the following processing is performed according to the flowchart shown in FIG. 14 .

First, at S600 , the parameters of the foot orbit of the current gait are set in such a way that the foot position and posture orbit of the current gait is connected to the foot position and posture orbit of the fixed gait.

The invention of the present application is characterized in that, in order to alleviate the impact of the robot when it lands, and to avoid slipping or spinning of the soles of the feet, so that the robot can walk or run stably, the trajectory parameters of the feet 22 are set in this way, but this will be discussed in It will be explained later.

Next, proceed to S602, and determine the reference upper body posture trajectory parameters of this gait in the same way as the first rotational gait and the second rotational gait of the fixed gait. Wherein, the reference upper body posture orbit of this gait is continuously connected with the reference upper body posture orbit of the fixed gait (the reference upper body posture angle and angular velocity at the terminal of this gait are respectively related to the fixed The initial reference upper body posture angle and angular velocity of the gait are consistent) to set the parameters. In addition, in the present embodiment, the reference upper body posture is a fixed vertical posture in any one of the current gait and the fixed gait.

Then, proceed to S604, and determine the arm posture trajectory parameters of this gait in the same way as the first rotational gait and the second rotational gait of the fixed gait. Wherein, the parameters are set in such a way that the arm posture track of the current gait is continuously connected with the arm posture track of the fixed gait. In addition, the arm posture trajectory parameters determined here are the same as the determination of the fixed gait parameters (S104 in Figure 12), except that the angular momentum of the two arms around the vertical axis (or the upper body trunk axis) changes. The motion parameter is a parameter specifying the trajectory of the center-of-gravity position of the two arms.

Thereafter, the process proceeds to S606, and the trajectory of the vertical component of the ground reaction force specified by this parameter is set to be a substantially continuous trajectory (with no step-like jump in value) as shown in the above-mentioned FIG. 9 .

Wherein, the parameters of the vertical component orbit of the ground reaction force are determined in such a manner that any one of the vertical position velocity of the overall center of gravity and the vertical component orbit of the ground reaction force of this gait is continuously connected to the fixed gait.

Afterwards, enter S608, and set the ground reaction force horizontal component allowable range [Fxmin, Fxmax] in the same way as the first rotation gait and the second rotation gait of the fixed gait (more specifically, the ground reaction force parameter of the model for the tolerance range of the horizontal component). For example, set the allowable range of the horizontal component of the ground reaction force according to the pattern shown in FIG. 15 . In the present embodiment, the allowable range of the horizontal component of the ground reaction force is set according to the formula 12 based on the vertical component mode of the ground reaction force determined in S606 above.

Then, proceed to S610, and set the ZMP of this gait as shown in FIG. Orbit (specifically, specify the parameters of the ZMP orbit, the time and position of the inflection point of the orbit). Wherein, the parameters are set in such a way that the ZMP orbit of the current gait is continuously connected with the ZMP orbit of the fixed gait. That is, the ZMP trajectory parameters are determined in such a way that the position of the ZMP at the end of this gait is consistent with the initial ZMP position of the fixed gait. In this case, in the running gait, the time and position of the inflection point of the ZMP trajectory in the single-leg stance phase can be set in the same way as the ZMP trajectory parameter setting method for the aforementioned fixed gait. Furthermore, the ZMP orbit parameter may be set so that the target ZMP orbit in the air phase changes continuously in a straight line from the start of the air phase to the initial ZMP position of the fixed gait.

In addition, the ZMP trajectory parameter of this gait determined at S610 is only temporarily determined, and will be modified as described later. Therefore, the ZMP trajectory of the current gait set as described above is hereinafter referred to as the temporary target ZMP trajectory of the current gait.

Returning to the description of FIG. 11 , after performing the processing shown in S026 (determination processing of the gait parameters of this gait), then enter S028 to modify the gait parameters (ZMP trajectory parameters) of this gait. In this process, the parameters of the ZMP trajectory are modified such that the upper body position posture trajectory is connected to or close to a fixed gait.

Returning to Fig. 11, after modifying the gait parameters of this time in S028 as mentioned above, or when the judgment result of S016 is NO, enter into S030, and determine the current gait parameters according to the modified gait parameters of this time. Instantaneous value of gait.

Then enter S032, and determine the arm movement for releasing the rotational force (the vertical component of the ground reaction force moment generated around the target ZMP is approximately zero through the movement of the robot 1 other than the arm). Specifically, the vertical component trajectory of the ground reaction force moment of the target ZMP without swinging the arm is obtained (strictly speaking, when the gait is generated without swinging the arm, the resultant force of the robot's gravity and inertial force acts on the target ZMP The sign of each instantaneous value of the moment vertical component orbit is flipped). That is, the instantaneous value is obtained by using the vertical component of the ground reaction force moment around the target ZMP (instantaneous value), and the vertical component of the ground reaction force moment around the target ZMP (instantaneous value) is related to the gait generated by the processing of S030. Instantaneous equalization of motion (this does not include arm swing motion). Then, by dividing the obtained instantaneous value by the equivalent moment of inertia of the arm swing motion, the angular acceleration of the arm swing motion required to release the rotational force is obtained. In addition, as a supplement, when the swing of the arm is too large, a value larger than the equivalent moment of inertia may be used as the divisor.

Then, this angular acceleration is integrated twice, and the second integrated value is passed through a low-cut filter for preventing the integrated value from being too large, and the obtained angle is used as an arm swing angle. However, in the arm swing motion, the left and right arms are swung forward and backward in opposite directions so that the positions of the centers of gravity of both arms do not change. In addition, the arm swing motion for releasing the rotational force may also be generated in advance in a fixed gait, and the arm swing motion in the current gait may be determined in a manner of being connected with the arm swing motion in the fixed gait.

Then proceed to S034, increase the time t for gait generation by Δt, return to S014, and continue the above-mentioned gait generation.

The above is the target gait generation process of the gait generation device 100 .

The operation of the device according to this embodiment will be further described with reference to FIG. 4 . In the gait generating device 100 , the target gait is generated as described above. The generated target upper body position posture (orbit) and target arm posture (orbit) in the target gait are output to the robot geometric model (inverse kinematics calculation unit) 102 .

In addition, target foot position posture (trajectory), target ZMP trajectory (target total ground reaction force center point trajectory), and target total ground reaction force (trajectory) (target ground reaction force horizontal component and target ground reaction force vertical component) is output to the composite adaptive action determination unit 104 and also output to the target ground reaction force distributor 106 . Then, the ground reaction force is distributed to the feet 22R and 22L by the target ground reaction force distributor 106 to determine the target ground reaction force center point of each foot and the target ground reaction force of each foot. The determined center point of the target ground reaction force of each foot and the target ground reaction force of each foot are output to the composite adaptive motion determination unit 104 .

The modified target foot position posture (trajectory) with mechanism deformation compensation is output from the composite adaptive motion determination unit 104 to the robot geometry model 102 . Once the robot geometry model 102 is input with the target upper body position and posture (orbit) and the modified target foot position and posture (orbit) with mechanism deformation compensation, the leg body 2, 2 that satisfies the two postures (orbit) is calculated. The joint displacement commands (values) of the 12 joints (10R(L) etc.) are sent to the displacement controller 108 . The displacement controller 108 controls the displacements of the 12 joints of the robot 1 following the joint displacement commands (values) calculated by the robot geometry model 102 as target values. In addition, the robot geometry model 102 also calculates the displacement designation (value) of the arm joint satisfying the target arm posture, and sends it to the displacement controller 108 . The displacement controller 108 controls the displacements of the 12 joints of the arm of the robot 1 following the joint displacement commands (values) calculated by the robot geometry model 102 as target values.

The ground reaction force generated on the robot 1 (specifically, the actual ground reaction force of each foot) is detected by the 6-axis force sensor 34 . The detected value is sent to the composite adaptive action determination unit 104 . In addition, the posture tilt deviations θerrx and θerry generated on the robot 1 (specifically refer to the deviation of the actual posture angle relative to the target upper body posture angle, the posture angle deviation in the left-right direction (around the X axis) is θerrx, and the front-back direction (around The posture angle deviation (θerry) on the Y axis) is detected by the inclination sensor 36 , and the detected value is sent to the posture stabilization control computing unit 112 . In this posture stabilization control calculation unit 112, the compensation total ground reaction force moment around the target total ground reaction force center point (target ZMP) for returning the upper body posture angle of the robot 1 to the target upper body posture angle is calculated. , and send it to the composite adaptive action determination unit 104. The composite adaptive action determination unit 104 modifies the target ground reaction force according to the input value. Specifically, the target ground reaction force is modified such that the compensating total ground reaction force moment acts around the target total ground reaction force center point (target ZMP).

The composite adaptive action determination unit 104 determines the modified target foot with mechanism deformation compensation if the actual robot state and ground reaction force calculated according to the sensor detection value and the like are consistent with the modified target ground reaction force. Position pose (orbit). However, since it is practically impossible to make all the states coincide with the target, a compromise relationship is taken among them to make them as consistent as possible. That is, control is performed by weighting the control deviation with respect to each target so that the weighted average of the control deviation (or the square of the control deviation) may become minimum. Accordingly, the actual foot position posture and the total ground reaction force are controlled so as to substantially follow the target foot position posture and the target total ground reaction force.

In addition, the gist of the present invention lies in the gait generation of the robot 1 in the gait generation device 100, because the configuration and actions of the above-mentioned compound adaptive motion determination unit 104 and the like have been described in detail in Japanese Patent Laying-Open No. 10, which was previously filed by the applicant. -277969 bulletin, etc., so the description stops here.

Next, generation of the walking gait of the robot 1 will be described. In addition, here, the walking gait does not have an air phase, but a gait in which the one-leg stance phase and the two-leg stance phase are repeated alternately.

When generating a walking gait, the following processes are performed in S106 and S606. That is, the phase and amplitude of the upper body vertical position trajectory (the upper body vertical position trajectory using the method for determining the height of the upper body disclosed in JP-A No. 10-86080 proposed by the present applicant) should be satisfied as much as possible. The characteristic quantity is to determine the vertical component track of the ground reaction force in this way, wherein the vertical position track of the upper body is based on the geometric conditions (geometric constraints) related to the displacement of each leg joint at least according to whether the bending angle of the knee is appropriate or not. conditions) are determined.

Accordingly, the main part of the gait generation algorithm can be shared between running and walking, and it is also possible to switch to running during walking or to walking during running.

The processing is described with reference to FIG. 17. First, at S1300, using the upper body height determination method previously applied by the applicant and disclosed in Japanese Patent Laid-Open No. 10-86080, etc., to obtain a height that satisfies at least 2 joints of each leg. The upper body vertical position track relative to the specified geometrical constraints. Hereafter, it is called the reference upper body vertical position orbit. More specifically, first, based on the foot orbital parameters and the target ZMP orbital parameters determined according to the required parameters, the upper body Horizontal position track. In addition, in this case, the vertical component of the ground reaction force is consistent with the self-weight of the robot 1, and the vertical position of the upper body is a predetermined value, and the horizontal component of the ground reaction force moment around the target ZMP is 0, In this way, the horizontal position track of the upper body is determined. In addition, the upper body posture trajectory at this time may be, for example, a trajectory of a certain posture (vertical posture, etc.).

Next, using the method for determining the height of the upper body proposed by the applicant in the previous application (JP-A-10-86080. More specifically, the method in FIG. The orbit and the upper body horizontal position orbit and upper body posture orbit determined as above are used to calculate the upper body vertical position orbit as the reference upper body vertical position orbit.

Then enter S1302, in order to determine the vertical component orbit of the ground reaction force, and calculate (extract) the characteristic quantities such as the amplitude and phase of the vertical position orbit of the reference upper body, wherein the vertical component orbit of the ground reaction force can be generated with the reference upper body The vertical position trajectory is as similar as possible to the vertical position trajectory of the upper body of the target. For example, the amplitude (the difference between the minimum value and the maximum value) of the reference upper body vertical position trajectory is calculated as the feature quantity.

Then enter S1304, so that the upper body vertical position track generated according to the ground reaction force vertical component track parameters can satisfy the characteristic quantity as much as possible (become a pattern similar to the reference upper body vertical position track as much as possible) ), so as to determine the orbital parameters of the vertical component of the ground reaction force (at the time of the inflection point and the value of the vertical component of the ground reaction force). More specifically, in the case of the walking gait, the first rotational gait and the second rotational gait of the fixed gait, and the vertical component trajectory of the ground reaction force of this gait are set as an example, as shown in Figure 16 of broken lines. That is, in the two-leg support period, the increase side of the vertical component of the ground reaction force is set as a convex (convex) trapezoidal shape, and in the single-leg support period, the vertical component of the ground reaction force is set as Convex (convex) trapezoidal shape. Moreover, the vertical component trajectory of the ground reaction force is integrated twice from the initial stage of the gait (the beginning moment of the two-leg support period) to the end point (the end moment of the single-leg support period), and the result obtained by the two integration The difference between the maximum value and the minimum value of the upper body vertical position track corresponding to the vertical position track of the overall center of gravity of the robot 1 is consistent with the characteristic quantity, so as to determine the vertical component track parameter of the ground reaction force, such as the ground The heights C1 and C2 of the two trapezoids of the vertical component trajectory of the reaction force (in this example, the moment of the inflection point of the vertical component trajectory of the ground reaction force is determined according to the required parameters related to the gait cycle).

However, the parameters of the vertical component trajectory of the ground reaction force for a fixed gait are also determined by satisfying the following conditions as mentioned above.

(Condition): The average value of the ground reaction force over the entire period of the fixed gait (periods of both the first rotation gait and the second rotation gait) in the vertical component orbit is equal to the self-weight of the robot. That is, the average value of the vertical component of the ground reaction force is equal to and opposite to the gravity acting on the robot.

In addition, the parameters of the vertical component track of the ground reaction force of this gait are determined in such a way that the track of the vertical position of the upper body (overall center of gravity) is continuously connected or close to the fixed gait as mentioned above.

From the above, the target ground reaction force vertical component trajectory (parameters specifying the vertical component trajectory) in the walking gait is determined. The gait generation processing other than the determination processing of the target ground reaction force vertical component trajectory described above may be the same as the embodiment related to the running gait described above.

As previously mentioned, the invention of the present case is characterized in that, in order to alleviate the impact of the robot when it lands, and avoid slipping or spinning of the soles of the feet so that the robot can walk or run stably, the trajectory parameters of the feet 22 are set in this way (refer to S600 ), which will be explained below.

The posture of the left foot 22L in the running gait of FIG. 5 seen from the lateral direction changes as shown in FIG. 6 . Thereupon, the distance d between the left foot 22 (or the sole) and the ground changes as shown in Figure 7, and the inclination angle θ of the foot 22 (or the sole) changes as shown in Figure 8 with respect to the ground, so that the foot is controlled The rotation action of the part 22 relative to the leg body 2. Specifically, first, at the beginning of the lift-off period (t=t1-t2), the left leg body 2L lands on the front end (toe) of the left foot 22L, and in this state, the left foot 22L is positioned relative to The inclination angle of the ground increases toward the positive (+) side, so that the rotation of the left foot 22L relative to the left leg body 2L is controlled, wherein the inclination angle of the left foot 22L relative to the ground is based on the foot joint 18L (or foot joint 18L). joints 18L and 20L), and the positive (+) side refers to the side where the rear end (heel) of the left foot 22L is farther from the ground than the front end (toe). In addition, the inclination angle ? After changing toward the negative (-) side in which the front end (toe) is higher than the rear end (heel) relative to the ground, it gradually approaches 0, so that the rotation of the left foot 22L relative to the left leg 2L is controlled. . The basic horizontal plane of the robot 1 or the inclination angle of the ground relative to a reference plane such as a horizontal plane is measured based on the inclination angle of the upper body 24 relative to the horizontal plane or based on image analysis of the ground captured by the camera 92, wherein the upper body The inclination angle of 24 relative to the horizontal plane is corresponding to the output of inclination sensor 36 when one or more legs 2 are on the ground.

The robot 100 of the present invention controls the left foot in such a way that the angle of inclination θ of the left foot 22L relative to the ground becomes 0 at the moment when it transitions from the aerial stage to the landing stage (starting moment of the landing stage) t=t7 at the latest. Part 22L rotates with respect to the left leg body 2L.

Alternatively, the inclination angle θ may be defined as a function θ(d) of the distance d. As mentioned above, from the middle moment of the leg body 2's lift-off period to the beginning moment of the landing period, the inclination angle θ of the foot portion 22 relative to the ground is gradually approaching 0, so as to control the angle of the foot portion 22 relative to the leg body 2. In this case (refer to FIG. 6 and FIG. 8 ), θ(d=0)=0, and |δθ/δd|>0 before transitioning from the ground-leaving phase to the landing phase. In addition, the angle θ may be controlled to be 0 during the air phase, and the angle θ may be maintained at 0 until the transition to the landing phase.

According to the robot 1 that exerts the above-mentioned function, the inclination angle θ of the foot portion 22 relative to the ground is gradually approached to 0 from the middle time of the ground-off phase of the foot portion 22 (or sole) of the leg body 2 to the start time of the ground-down phase. In this way, the rotation movement of the foot part 22 relative to the leg body 2 is controlled (see FIG. 5 , FIG. 6 and FIG. 8 ). Accordingly, since the ground area of the feet 22 (or soles) of the leg body 2 after transitioning from the lift-off period to the landing period is large, the impact when it lands can be widely distributed to the soles of the feet, thereby mitigating the impact of the robot. 1 was hit. In addition, since the friction between the feet 22 and the ground is large, even if the moving speed and the angular velocity of the yaw of the robot 1 before the leg body 2 touches the ground are large, this friction can prevent the robot 1 from transitioning to the landing stage. Slip and spin. Therefore, the robot 1 of the present invention can reduce the impact when the feet 22 of the legs 2 land on the ground, and can also avoid slipping and spinning at the feet 22, so that they can run stably.

In addition, the rotation motion of the foot 22 relative to the leg body 2 is controlled so that the front end (toe) of the foot 22 kicks the ground (refer to the trajectory of the foot position and posture at times t1 to t2 in FIG. 6 ). According to this, the propulsion force of the robot 1 is enhanced. On the other hand, since the feet 22 of the robot 1 can be prevented from slipping and spinning when landing as described above, the robot 1 can move at a high speed while stabilizing its motion. .

In addition, from the start time of the lift-off period to the start time of the touch-down period, the foot 22 changes from a heel upturned posture with respect to the ground (see the foot position posture trajectory at times t2 to t4 in FIG. 6 ) to the toe position. Upturned posture (refer to the foot position posture track at time t5-t6 in Fig. 6), and then close to the posture parallel to the ground, so that the ground area of the foot 22 can be ensured as mentioned above to prevent the robot from 1 degree of slipping and spinning.

In addition, even when walking in the air without both legs 2 leaving the ground simultaneously (see FIG. That is, when the robot 1 is walking, as described above regarding running, for example, the inclination angle θ of the feet 22 relative to the ground can be gradually approached from the middle moment of the leg body 2's lift-off period to the start moment of the landing period. 0, the rotation of the foot 22 relative to the leg body 2 is controlled in this way.

Accordingly, because the foot 22 (or sole) of the leg body 2 after the transition from the lift-off period to the landing period has a relatively large ground area, the impact when it lands can be widely distributed to the soles of the feet, thereby reducing the impact on the ground. Impact on robot 1. In addition, because the friction between the feet 22 and the ground is relatively large, even if the moving speed and the angular velocity of the yaw of the robot 1 before the feet of the legs 2 touch the ground are large, the friction can prevent the transition to the ground after the grounding period. Slipping and spinning of robot 1. Therefore, the robot 1 of the present invention can walk stably by avoiding slipping and spinning of the legs 22 while mitigating the impact when landing.

Claims (18)

1. leg type mobile robot, it is linked to a plurality of leg bodies of matrix by driving, come on one side to change repeatedly ground reaction force act on phase that lands on any one foot in a plurality of leg bodies, with the aerial phase that on the foot of all leg bodies, does not all have the ground reaction force effect, move on one side, it is characterized in that

Be converted to from the aerial phase when landing the phase, make the foot of the predetermined leg body that lands change gradually, and the tread surface of foot be parallel to the ground when this leg body lands, so the driving leg body with respect to the angle of inclination on ground.

2. leg type mobile robot according to claim 1 is characterized in that,

Before the leg body will be liftoff, under the state that the leading section of the foot of this leg body still lands, make the rearward end of this foot leave gradually from ground, so drive this leg body.

3. leg type mobile robot according to claim 1 is characterized in that,

Be carved into the finish time during from liftoff phase of leg body middle, make the foot leading section, so drive this leg body from being that the higher state of benchmark changes to equal height gradually with the foot rearward end.

4. leg type mobile robot according to claim 1 is characterized in that,

From zero hour of liftoff phase of leg body to the centre constantly, make the foot leading section from being that the lower state of benchmark changes to equal height gradually with the foot rearward end, become higher state afterwards gradually, so drive this leg body.

5. leg type mobile robot, a plurality of leg bodies that it has the upper body and extends setting from the upper body downwards, and by with respect to the liftoff of each rotary foot of leg body and the action of each the leg body that accompanies of landing move, it is characterized in that, be provided with:

Mechanism is measured at the foot angle of inclination, and it measures the angle of inclination of foot with respect to ground; And

The foot action controlling organization, it is carved into the zero hour of the phase of landing during from liftoff phase of leg body middle, feasible foot by foot's angle of inclination mensuration this leg body that mechanism measured approaches 0 gradually with respect to the angle of inclination on ground, so controls the rotational action of foot with respect to this leg body.

6. leg type mobile robot according to claim 5 is characterized in that,

The foot action controlling organization is controlled the rotational action of foot with respect to the leg body in following mode, promptly, before the near liftoff phase, under the state that the leg body still lands with the leading section of foot, make that measuring this foot that mechanism measured by the foot angle of inclination increases than the positive side of leading section away from ground towards the rearward end of this foot with respect to the angle of inclination on ground.

7. leg type mobile robot according to claim 5 is characterized in that,

The foot action controlling organization is controlled the rotational action of foot with respect to the leg body in following mode, promptly, be carved into the zero hour of the phase of landing during from liftoff phase of leg body middle, make that measuring foot that mechanism measured by the foot angle of inclination gradually reduces to 0 than rearward end away from the angle of the minus side on ground from the leading section of foot with respect to the angle of inclination on ground.

8. leg type mobile robot according to claim 6 is characterized in that,

The foot action controlling organization is controlled the rotational action of foot with respect to the leg body in following mode, promptly, from the zero hour of the zero hour to phase of landing of liftoff phase of leg body, make that measuring the foot that mechanism measured by the foot angle of inclination reduces again after positive side increases gradually gradually with respect to the angle of inclination on ground, gradually increase than the minus side of rearward end towards the leading section of foot again then, reduce to 0 gradually afterwards away from ground.

9. leg type mobile robot according to claim 5 is characterized in that,

Described leg type mobile robot is accompanied by the liftoff aerial phase of all leg bodies and moves.

10. control program, it is the program that the function that will be used to control following leg type mobile robot offers the computer of lift-launch in this robot, this leg type mobile robot is linked to a plurality of leg bodies of matrix by driving, come on one side to change repeatedly ground reaction force act on phase that lands on any one foot in a plurality of leg bodies, with the aerial phase that on the foot of all leg bodies, does not all have the ground reaction force effect, move on one side, described control program is characterised in that

Following function is offered the computer of lift-launch in described robot, this function is: be converted to from the aerial phase when landing the phase, the foot of the feasible predetermined leg body that lands changes gradually with respect to the angle of inclination on ground, and the tread surface of foot is parallel to the ground when this leg body lands, and so controls the function of action of the leg body of described robot.

11. control program according to claim 10 is characterized in that,

Following function is offered the computer of lift-launch in described robot, this function is: before the leg body will be liftoff, under the state that this leg body still lands with the leading section of its foot, make the rearward end of this foot leave gradually, so control the function of action of the leg body of described robot from ground.

12. control program according to claim 10 is characterized in that,

Following function is offered the computer of lift-launch in described robot, this function is: be carved into the finish time during from liftoff phase of leg body middle, make the foot leading section from being that the higher state of benchmark changes to equal height gradually with the foot rearward end, so control the function of action of the leg body of described robot.

13. control program according to claim 10 is characterized in that,

Following function is offered the computer of lift-launch in described robot, this function is: from zero hour of liftoff phase of leg body to the centre constantly, make the foot leading section from being that the lower state of benchmark changes to equal height gradually with the foot rearward end, gradually become higher state afterwards, so control the function of action of the leg body of described robot.

14. control program, it is the program that the function that will be used to control following leg type mobile robot offers the computer of lift-launch in this robot, promptly, a plurality of leg bodies that this leg type mobile robot has the upper body and extends setting from the upper body downwards, and by with respect to the liftoff of each rotary foot of leg body and the action of each the leg body that accompanies of landing move, described control program is characterised in that

Following foot's angle of inclination measurement function and foot action control function are offered the computer of lift-launch in described robot,

Described foot angle of inclination measurement function is: measure the angle of inclination of foot with respect to ground;

Described foot action control function is: the zero hour that is carved into the phase of landing during from liftoff phase of leg body middle, make and to utilize the foot of this leg body that foot's angle of inclination measurement function measured to approach 0 gradually, so control the rotational action of foot with respect to this leg body with respect to the angle of inclination on ground.

15. control program according to claim 14 is characterized in that,

To offer the computer of lift-launch in described robot as the following function that foot action is controlled function, this function is: before the near liftoff phase, under the state that the leg body still lands with the leading section of foot, make and to utilize this foot that foot's angle of inclination measurement function measured to increase than the positive side of leading section towards the rearward end of this foot, so control the function of foot with respect to the rotational action of this leg body away from ground with respect to the angle of inclination on ground.

16. control program according to claim 14 is characterized in that,

To offer the computer of lift-launch in described robot as the following function that foot action is controlled function, this function is: the zero hour that is carved into the phase of landing during from liftoff phase of leg body middle, make and to utilize foot that foot's angle of inclination measurement function measured gradually to reduce to 0 than rearward end away from the angle of the minus side on ground from the leading section of foot, so control the function of foot with respect to the rotational action of this leg body with respect to the angle of inclination on ground.

17. control program according to claim 15 is characterized in that,

To offer the computer of lift-launch in described robot as the following function that foot action is controlled function, this function is: from the zero hour of the zero hour to phase of landing of liftoff phase of leg body, make the foot that utilizes foot's angle of inclination measurement function to be measured after positive side increases gradually, reduce gradually again with respect to the angle of inclination on ground, gradually increase than the minus side of rearward end towards the leading section of foot again then away from ground, gradually reduce to 0 afterwards, so control the function of foot with respect to the rotational action of this leg body.

18. control program according to claim 10 is characterized in that,

Following function is offered the computer of lift-launch in described robot, and this function is: make described robot be accompanied by the liftoff aerial phase of all leg bodies and move, so control the function of action of the leg body of described robot.

Images